March 2026 arXiv papers — page 110

Md. Asraful Haque, Aasar Mehdi, Maaz Mahboob, Tamkeen Fatima

Large Language Models (LLMs) have achieved unprecedented fluency but remain susceptible to "hallucinations" - the generation of factually incorrect or ungrounded content. This limitation is particularly critical in high-stakes domains where reliability is paramount. We propose a domain-grounded tiered retrieval and verification architecture designed to syste

Lucien Cousin, Pietro Miotti, Bruno Marco-Dufort, Igor V. Pivkin

Adding dynamic bonds in polymer networks enables reprocessing and recycling; however the full impact of reversible bonds on dynamic network mechanics remains unclear. We build model dynamic networks and observe substantial deviations from classic theory. We rationalize these findings by considering that bond exchange enables the networks to rearrange and ado

A. Nourou Issa

In this paper the notion of a quadratic (left) Bol algebra is discussed. Several examples of quadratic Bol algebras are given and it is observed that the only two-dimensional quadratic real Bol algebras are quadratic Lie triple systems. Dual representations of Bol algebras are investigated with a particular emphasis on coadjoint representations for quadratic

Oleg Pikhurko, Kohki Sakamoto

For $n \ge 2$, Gamburd, Jakobson, and Sarnak [J. Eur. Math. Soc. 1, 51-85 (1999)] conjectured that almost every $n$-tuple in $\mathrm{SU}(2)$ has a spectral gap. Toward this conjecture, Fisher [Int. Math. Res. Not. (2006)] established a zero-one law for $n \ge 3$, but obtained only a partial result for $n=2$. In this paper, we prove that the zero-one law als

Damian Brzozowski, Sander R. Hønnås, Egil Y. Tokle, Jørgen A. Arnesen

Metal chalcogenides are promising layered topological materials, yet their electronic performance is often limited by parasitic bulk conduction arising from defects that introduce excess carriers and shift the Fermi level out of the topological regime. Controlling early-stage growth and defect formation is therefore essential for suppressing bulk transport a

Marijn Ruiter, Miguel Aguiar, Jake Rap, Karl H. Johansson

We propose RHYME-XT, an operator-learning framework for surrogate modeling of spatiotemporal control systems governed by input-affine nonlinear partial integro-differential equations (PIDEs) with localized rhythmic behavior. RHYME-XT uses a Galerkin projection to approximate the infinite-dimensional PIDE on a learned finite-dimensional subspace with spatial

Quang Nguyen, Ronald Yurko

In sports analytics, player tracking data have driven significant advancements in the task of player evaluation. We present a novel generative framework for evaluating the observed frame-by-frame player positioning against a distribution of hypothetical alternatives. We illustrate our approach by modeling the within-play movement of an individual ball carrie

Nuria Senar, Stavros Makrodimitris, Michel H. Hof, Cornelis Verhoef

Detection of minimal residual disease (MRD) in cancer patients after surgery can provide an early marker for disease recurrence and guide subsequent treatment decisions. Accurate and sensitive estimation of tumour burden after cancer surgery may be obtained through liq- uid biopsies, measuring circulating tumour DNA (ctDNA) using, for example, mutation-based

Fangyijie Wang, Guénolé Silvestre, Kathleen M. Curran

Congenital heart disease (CHD) screening from fetal echocardiography requires accurate analysis of multiple standard cardiac views, yet developing reliable artificial intelligence models remains challenging due to limited annotations and variable image quality. In this work, we propose FM-DACL, a semi-supervised Dual Agreement Consistency Learning framework

Mark Braverman, Jingyi Liu, Eric Xue, Chenghan Zhou

We study one-sided matchings with endowments in the absence of money. It is well-known that a competitive equilibrium may not always exist and that the strong core may be empty in this setting [Hylland and Zeckhauser, 1979]. We propose a generalization of competitive equilibria that associates each item with a multi-dimensional price. We show that this solut

J. -R. Chazottes, P. Collet, F. Redig

We develop a transport-entropy framework for Gaussian concentration inequalities on the infinite product space $S^{\mathbb Z^d}$, where $S$ is a finite set, in which sensitivity is measured by the $\ell^2$-norm of local oscillations. We show that the associated transportation costs cannot be induced by any metric or cost function on the configuration space,

Byron Dowling, Jacob Piland, Eleanor Frederick, Christopher Sweet

Human perceptual priors have shown promise in saliency-guided deep learning training, particularly in the domain of iris presentation attack detection (PAD). Common saliency approaches include hand annotations obtained via mouse clicks and eye gaze heatmaps derived from eye tracking data. However, the most effective form of human saliency for raising general

Sebastian Bechtel, Esmée Theewis

We establish existence of probabilistically strong solutions and pathwise uniqueness for a class of quasilinear stochastic evolution equations on bounded domains. Our results combine recent weak existence results for quasilinear stochastic evolution equations in an $L^p$-setting (with $p > 2$) with Yamada--Watanabe theory. To establish pathwise uniqueness, w

Seishiro Ono, Yanbai Zhang, Hoi Chun Po

Accurate contraction of tensor networks beyond one dimension is essential in various fields including quantum many-body physics. Existing approaches typically rely on approximate contraction schemes and do not provide certified error bars. We introduce a numerical bootstrap framework which casts the problem of tensor-network contractions into a convex optimi

William Thorossian

Modern seismic and volcanic monitoring is increasingly shaped by continuous, multi-sensor observations and by the need to extract actionable information from nonstationary, noisy wavefields. In this context, machine learning has moved from a research curiosity to a practical ingredient of processing chains for detection, phase picking, classification, denois

Manish Kumar, Deng-Yuan Li, Zhangyu Yuan, Ying Wang

Quantum spin rings represent an intriguing platform for studying unconventional magnetic order and exotic quantum phases, and they are also promising materials for emerging quantum technologies. Conventional spin systems consist of a set of weakly interacting localized spins that are well described by the Heisenberg spin models. Here, we demonstrate that str

William Balderrama, Jack Morgan Davies, Sil Linskens

We introduce generalizations of global equivariant spectra which encode globally equivariant cohomology theories equipped with additional transfers, such as the deflation maps present in equivariant topological $K$-theory. We call these $\mathcal{Q}$-ambidextrous global spectra, where $\mathcal{Q}$ is a parameter encoding which additional transfers one allow

Oussama Bensaid, Anthony Genevois, Romain Tessera

We study coarse separation in one-ended hyperbolic groups from a quantitative point of view, focusing on the volume growth of separating subsets. We prove that a one-ended hyperbolic group that is not virtually a surface group is coarsely separable by a subset of subexponential growth if and only if it splits over a virtually cyclic subgroup. To do so, we sh

Xitong Chen, Yifeng Pan, Min Li, Xiaotian Ding

Large-scale, high-quality multimodal demonstrations are essential for robot learning of contact-rich dexterous manipulation. While human-centric data collection systems lower the barrier to scaling, they struggle to capture the tactile information during physical interactions. Motivated by this, we present DexViTac, a portable, human-centric data collection

Zhou Fang, Jiaqi Wang, Yi Zhou, Qiongfeng Shi

Recent Vision-Language-Action (VLA) models equipped with Flow Matching (FM) action heads achieve state-of-the-art performance in complex robot manipulation. However, the multi-step iterative ODE solving required by FM introduces inference latency that precludes responsive physical control. While current acceleration efforts optimize the Vision-Language Model

Stefan Schwede

We prove a splitting result in global equivariant homotopy theory that is a simultaneous refinement of the Segal--Becker splitting and its `Real' and equivariant generalizations, and of the explicit Brauer induction of Boltje and Symonds. We show that the morphism of ultra-commutative Real-global ring spectra from $\Sigma^\infty_+ B_{\text{gl}}U(1)$ to the R

Paolo Marcandelli, Stefano Mariani, Martina Siena, Stefano Markidis

Fourier representations play a central role in operator learning methods for partial differential equations and are increasingly being explored in quantum machine learning architectures. The classical fast Fourier transform (FFT), particularly in its Cooley--Tukey decomposition, exhibits a structure that naturally matches continuous-variable quantum circuits

Sudaice Kazibwe, Bishnu Karki, Wencheng Lu, Zhongxin Liang

AgSbTe2 is a well-known thermoelectric material with a high Seebeck coefficient and intrinsically low thermal conductivity, but its behavior under pressure remains largely unexplored. Here we report a systematic investigation of the structural, electronic, and transport properties of non-stoichiometric AgSbTe2 under high pressure. At ambient pressure, the ma

Anwai Archit, Constantin Pape

Cell segmentation is a fundamental task in microscopy image analysis. Several foundation models for cell segmentation have been introduced, virtually all of them are extensions of Segment Anything Model (SAM), improving it for microscopy data. Recently, SAM2 and SAM3 have been published, further improving and extending the capabilities of general-purpose seg

G. Tartaglione, G. Zanfardino, F. Illuminati

We derive an exact uncertainty relation for arbitrary quantum states of finite-dimensional Hilbert spaces. For any given $k$-partition of a $d$-dimensional multipartite system, we introduce the total uncertainty as the sum of the uncertainties associated with all possible tensor products of local $\mathrm{SU}(N)$ observables, where each observable acts on th

Christopher J. Conselice, Edmund J. Copeland, Sergio Sevillano Muñoz

In this study we adapt a classical cosmology measurement, the volume or number density test, to a modern synthesis of observed galaxy evolution. We do this by using measured galaxy mass functions and the history of galaxy evolution through star formation and galaxy mergers, inspired by the latest results from deep extragalactic surveys. We develop a new fram

Chen Liyi, Wang Pengfei, Zhang Guowen, Ma Zhiyuan

Most instruction-driven 3D editing methods rely on 2D models to guide the explicit and iterative optimization of 3D representations. This paradigm, however, suffers from two primary drawbacks. First, it lacks a universal design of different 3D editing tasks because the explicit manipulation of 3D geometry necessitates task-dependent rules, e.g., 3D appearanc

Zhaochong An, Zirui Li, Mingqiao Ye, Feng Qiao

Video understanding aims to enable models to perceive, reason about, and interact with the dynamic visual world. In contrast to image understanding, video understanding inherently requires modeling temporal dynamics and evolving visual context, placing stronger demands on spatiotemporal reasoning and making it a foundational problem in computer vision. In th

Yao Wang, Xin Liu, Zhuochen Liu, Jiankang Chen

Existing human value datasets do not directly support value understanding in factual news: many are actor-agnostic, rely on isolated utterances or synthetic scenarios, and lack explicit event structure or value direction. We present \textbf{NEVU} (\textbf{N}ews \textbf{E}vent-centric \textbf{V}alue \textbf{U}nderstanding), a benchmark for \emph{actor-conditi

Petros Ellinas, Indrajit Chaudhuri, Johanna Vorwerk, Spyros Chatzivasileiadis

Physics-informed machine learning surrogates are increasingly explored to accelerate dynamic simulation of generators, converters, and other power grid components. The key question, however, is not only whether a surrogate matches a stand-alone component model on average, but whether it remains accurate after insertion into a differential-algebraic simulator

Yuhao Dong, Zuyan Liu, Shulin Tian, Yongming Rao

Large Language Models (LLMs) have achieved remarkable reliability and advanced capabilities through extended test-time reasoning. However, extending these capabilities to Multi-modal Large Language Models (MLLMs) remains a significant challenge due to a critical scarcity of high-quality, long-chain reasoning data and optimized training pipelines. To bridge t

Zunzhe Zhang, Runhan Huang, Yicheng Liu, Shaoting Zhu

Diffusion models and flow matching have become a cornerstone of robotic imitation learning, yet they suffer from a structural inefficiency where inference is often bound to a fixed integration schedule that is agnostic to state complexity. This paradigm forces the policy to expend the same computational budget on trivial motions as it does on complex tasks.

Bassam Adnan, Aviral Gupta, Sreemaee Akshathala, Karthik Vaidhyanathan

Benchmarks for large language models (LLMs) have progressed from snippet-level function generation to repository-level issue resolution, yet they overwhelmingly target implementation correctness. Software architecture tasks remain under-specified and difficult to compare across models, despite their central role in maintaining and evolving complex systems. W

Jefferson Hernandez, Swarnadeep Saha, Chenxi Whitehouse, Sanjeel Parekh

In this work, we probe the ability of a language model to demonstrate spatial reasoning from unstructured text, mimicking human capabilities and automating a process that benefits many downstream media applications. Concretely, we study the narrative-to-play task: inferring stage-play layouts (scenes, speaker positions, movements, and room types) from text t

Zhenhang Yuan, Shenghai Yuan, Lihua Xie

LLM agents often fail in closed-world embodied environments because actions must satisfy strict preconditions -- such as location, inventory, and container states -- and failure feedback is sparse. We identify two structurally coupled failure modes: (P1) invalid action generation and (P2) state drift, each amplifying the other in a degenerative cycle. We pre

Sofía Pinto, P. A. Orellana, Sergio Bravo

One-dimensional crystals serve as a versatile platform for engineering nontrivial states, which can be easily explored in transport configurations. In this work, we analyze the properties of a periodic structure composed of an H-shaped unit cell, which forms a periodic ladder-shaped system. Using tight-binding models, group-theoretical considerations, and st

Lintang Sutawika, Aditya Bharat Soni, Bharath Sriraam R R, Apurva Gandhi

A prerequisite for coding agents to perform tasks on large repositories is code localization - the identification of relevant files, classes, and functions to work on. While repository-level code localization has been performed using embedding-based retrieval approaches such as vector search, recent work has focused on developing agents to localize relevant

Qianlong Xiang, Miao Zhang, Haoyu Zhang, Kun Wang

Although text-to-image diffusion models exhibit remarkable generative power, concept erasure techniques are essential for their safe deployment to prevent the creation of harmful content. This has fostered a dynamic interplay between the development of erasure defenses and the adversarial probes designed to bypass them, and this co-evolution has progressivel

Xiaomeng Sui, Allard Mosk

Transmission matrices are valuable tools to describe and control light transport through scattering media. There are only a few cases where the transmission matrix can be compared to microscopic theories. Here we measure the polarization-complete transmission matrix of a single dielectric sphere using off-axis holography with angle scanning and reconstruct c

Ruize Ma, Yilei Jiang, Shilin Zhang, Zheng Ma

Multimodal Automated Program Repair (MAPR) extends traditional program repair by requiring models to jointly reason over source code, textual issue descriptions, and visual artifacts such as GUI screenshots. While recent LLM-based repair systems have shown promising results, existing approaches face several limitations: rigid workflow pipelines restrict expl

Xianhang Cheng, Yujian Zheng, Zhenyu Xie, Tingting Liao

In this work, we study the role of Massive Activations (MAs), which are rare, high-magnitude spikes confined to a few fixed hidden dimensions in video diffusion transformers (DiTs). We uncover a structured positional hierarchy: MA magnitudes peak at first-frame tokens and recur at the spatial boundary tokens of latent frames, with this pattern being most pro

Jing Qin, Muhao Chen

Tensegrity structures possess intrinsic geometric symmetries that govern their dynamic behavior. However, most existing physics-informed neural network (PINN) approaches for tensegrity dynamics do not explicitly exploit these symmetries, leading to high computational complexity and unstable optimization. In this work, we propose a symmetry-reduced physics-in

Jianwei Zhang

Background: Amid the opportunities and risks introduced by generative AI, learning research needs to envision how human minds and responsibilities should re-adapt as AI augments or automates various tasks and enters daily learning, knowledge work, and social life. Approach: Drawing on theories of learning, intelligence, and knowledge creation, this conceptua

Yanke Yu, Jin Li, Ying Sun, Ping Li

Understanding the internal functional organization of Large Language Models (LLMs) is crucial for improving their trustworthiness and performance. However, how LLMs organize different functions into modules remains highly unexplored. To bridge this gap, we formulate a functional module discovery problem and propose an Unsupervised LLM Cross-layer MOdule Disc

Aivo Olev, Tanel Alumäe

Large audio language models (LALMs) can answer questions about speech, music, and environmental sounds, yet their internal reasoning is largely opaque and difficult to validate. We describe TalTech's solution to the Agent Track of the Interspeech 2026 Audio Reasoning Challenge, in which systems are evaluated on reasoning process quality, specifically the fac

Md Mostafizer Rahman, Ariful Islam Shiplu, Yutaka Watanobe, Md Faizul Ibne Amin

Contextual embeddings generated by LLMs exhibit strong positional inductive biases, which can limit their ability to fully capture long-range, order-sensitive dependencies in highly structured source code. Consequently, how to further refine and enhance LLM embeddings for improved code understanding remains an open research question. To address this gap, we

David Millard, Cecilia Alm, Rashid Ali, Pengcheng Shi

Federated reinforcement learning typically aggregates value functions or policies by parameter averaging, which emphasizes expected return and can obscure statistical multimodality and tail behavior that matter in safety-critical settings. We formalize federated distributional reinforcement learning (FedDistRL), where clients parametrize quantile value funct

Émilie Charlier, Savinien Kreczman, Zuzana Masáková, Edita Pelantová

Alternate bases are a numeration system that generalizes the Rényi numeration system. It is common in this context to construct examples or counter-examples by specifying the expansions of $1$ in the desired system. While it is easy to show when a system with given expansions of $1$ exists in the Rényi case, the same is not true in the alternate case. In thi

Joel Bobadilla

This thesis investigates the magnetic, spectral, and transport properties of strongly correlated electronic systems, with a primary focus on the Hubbard model and its extensions relevant for real materials. Within the dynamical mean-field theory (DMFT) framework, different regimes of interaction strength, temperature, doping, and magnetic field are explored,

Ales Prokes, Tomas Mikulasek, Josef Vychodil, Radek Zavorka

The aim of this paper is to provide a comparison of channel characteristics for vehicle-to-vehicle (V2V) communication at 60 GHz and 80 GHz frequency bands in a high-mobility scenario where two vehicles pass each other in opposite directions. The study is based on measurements of the time-varying channel impulse response capturing the behavior of multi-path

Robin Ollive, Stéphane Louise

To process a problem with a Quantum Processing Unit (QPU), it must be transformed into a sequence of quantum operators, or gates. These operators are either packed into a query (i.e. quantum algorithm primitive) that encodes the problem, or used to construct the cost function for Variationnal Quantum Algorithm (VQA). Typical queries are the problem Hamiltoni

Muffy Calder, Marion Oswald, Elizabeth McClory-Tiarks, Michele Sevegnani

There is growing interest in the use of Large Language Models (LLMs) in policing, but there are potential risks. We have developed a practical approach to identifying risks, grounded in the policing and legal system of England and Wales. We identify 15 policing tasks that could be implemented using LLMs and 17 risks from their use, then illustrate with over

Corentin Royer, Debarun Bhattacharjya, Gaetano Rossiello, Andrea Giovannini

Multi-step reasoning improves the capabilities of large language models (LLMs) but increases the risk of errors propagating through intermediate steps. Process reward models (PRMs) mitigate this by scoring each step individually, enabling fine-grained supervision and improved reliability. Existing methods for training PRMs rely on costly human annotations or

Yonghui Zhang, Rui Zhu, Huili Liang, Guochao Zhao

Conventional polarization detection relies on external filters, which incur significant efficiency loss and polarization crosstalk, especially in the deep ultraviolet band where subwavelength nanofabrication is challenging. Here, we report that monoclinic \b{eta}-Ga2O3 exhibits intrinsic giant polarization dichroism, allowing near-ideal polarization photodet

Qiangqiang Wu, Tianyu Yang, Bo Fang, Jia Wan

Tracking Any Point (TAP) has emerged as a fundamental tool for video understanding. Current approaches adapt Vision Foundation Models (VFMs) like DINOv2 via offline finetuning or test-time optimization. However, these VFMs rely on static image pre-training, which is inherently sub-optimal for capturing dense temporal correspondence in videos. To address this

Dmitriy Rivkin, Parker Ewen, Lili Gao, Julian Ost

Recent video diffusion models achieve high-quality generation through recurrent frame processing where each frame generation depends on previous frames. However, this recurrent mechanism means that training such models in the pixel domain incurs prohibitive memory costs, as activations accumulate across the entire video sequence. This fundamental limitation

Antônio Junior Alves Caiado, Michael Hahsler

Transformer-based language models are widely deployed for reasoning, yet their behavior under inference-time stochasticity remains underexplored. While dropout is common during training, its inference-time effects via Monte Carlo sampling lack systematic evaluation across architectures, limiting understanding of model reliability in uncertainty-aware applica

Ziwei Xiang, Fanhu Zeng, Hongjian Fang, Rui-Qi Wang

Large Vision Language Models (LVLMs) have achieved remarkable success in a range of downstream tasks that require multimodal interaction, but their capabilities come with substantial computational and memory overhead, which hinders practical deployment. Among numerous acceleration techniques, post-training quantization is a popular and effective strategy for

Ruixiang Wang, Qingming Liu, Yueci Deng, Guiliang Liu

Video generative models are increasingly used as world models for robotics, where a model generates a future visual rollout conditioned on the current observation and task instruction, and an inverse dynamics model (IDM) converts the generated frames into executable robot actions. However, current video world models lack explicit executability constraints. A

Jing Wang, Jie Shen, Amar Sra, Qiaomin Xie

Phenotypic characterization is essential for understanding heterogeneity in chronic diseases and for guiding personalized interventions. Long COVID, a complex and persistent condition, yet its clinical subphenotypes remain poorly understood. In this work, we propose an LLM-augmented computational phenotyping framework ``Grace Cycle'' that iteratively integra

Bachana Beradze, Mikheil Tsitsishvili, Sergej Moroz

We investigate the phase diagram and critical properties of a one-dimensional $\mathbb{Z}_{2}$ lattice gauge theory describing an orthogonal metal, where spinless fermions and Ising spins are minimally coupled to a deconfined $\mathbb{Z}_{2}$ gauge field. Working at half-filling of fermions, we derive an exact gauge-invariant formulation that maps the model

Anthony Maocheia-Ricci, Edith Law

Value-based approaches such as Value Sensitive Design (VSD) enable technology designers to engage with and integrate human values in technology through a tripartite methodology of conceptual, empirical, and technical investigations. However, VSD contains pitfalls in both translating values to requirements and a lack of normative grounding, leading to adaptat

Colin Desmarais

In this work, recent results on the moments of balanced P\'olya urns are generalized to unbalanced urns, with the condition that the expected change in total activity at each step is constant. We also provide applications of our results to the degree distributions of random trees grown by uniform attachment with freezing and to the degree distribution of hoo

Tuowei Wang, Liyun Chu, Ruwen Fan, Ju Ren

The key-value (KV) cache has become the dominant contributor to memory consumption in large language model (LLM) inference. Although offloading KVCache from GPU high-bandwidth memory (HBM) to CPU DRAM alleviates device memory pressure, DRAM remains capacity-limited and costly for large, persistent workloads. Solid-state drives (SSDs) provide a cost-effective

Qihao Lin, Borui Chen, Yuping Zhou, Jianing Wu

The contour estimation of transparent fragments is very important for autonomous reassembly, especially in the fields of precision optical instrument repair, cultural relic restoration, and identification of other precious device broken accidents. Different from general intact transparent objects, the contour estimation of transparent fragments face greater

Siyuan Li, Chengyuan Liu, Wen-Hua Chen

This paper presents a unified decision-making framework that integrates Hybrid Markov Decision Processes (HMDPs) with Model Predictive Control (MPC), augmented by velocity-dependent safety margins and a prediction-aware hysteresis mechanism. Both the ego and surrounding vehicles are modeled as HMDPs, allowing discrete maneuver transition and kinematic evolut

Zahra Golsanamlou, Arseniy Baskakov, Robbert van de Kruijs, Silvester Houweling

In this work, we simulate the spectral emissivity of various stoichiometric crystal phases of Mo$_x$Si$_y$ compounds using density functional perturbation theory. The dielectric function, including electronic and ionic contributions, is calculated for each phase. We use the bulk properties obtained to simulate the optical absorption spectrum originating from

Jie Lei, Héctor Martínez, Adrián Castelló

The growing adoption of RISC-V in high-performance and scientific computing has increased the need for performance-portable code targeting the RISC-V Vector (RVV) extension. However, current compiler infrastructures provide limited end-to-end support for generating optimized RVV code from high-level representations to low-level implementations. In particular

Ranojit Barman, Masaaki Kimura, Yoshiki Chazono, Kazuki Yoshida

Background: Neutron-rich nucleus $^{44}$S lies in the region where traditional $N=28$ shell closure weakens, leading to the emergence of shape coexistence and large-amplitude collective motion (LACM). Understanding the nature and degree of shape mixing in this nucleus remains an important and fascinating problem. Purpose: We investigate the manifestation of

Yanpeng Zhou, Gang Li

Quantum spin liquids (QSLs) are exotic phases of matter characterized by long-range entanglement and the absence of magnetic order even at zero temperature. Here, we present a comprehensive theoretical study of the frustrated magnet Y$_3$Cu$_2$Sb$_3$O$_{14}$ to elucidate its electronic and magnetic properties. We uncover completely opposite crystal-field spl

Luca Hinkamp, Simon Klüttermann, Emmanuel Müller

In practice, machine learning methods commonly require anomaly detection (AD) to filter inputs or detect distributional shifts. Typically, this is implemented by running a separate AD model alongside the primary model. However, this separation ignores the fact that the primary model already encodes substantial information about the target distribution. In th

Zhong-Hua Zhang, Xi-Hu Lv, Xu-Guang Huang

We formulate a relativistic hydrodynamic theory for fluids with spin and intrinsic dilation charges. Using an entropy-current analysis, we derive constitutive relations featuring a bulk viscosity and a dilation conductivity governing the relaxation and diffusion of dilation charge. Linear mode analysis reveals a gapped dilation excitation and the freeze-out

Jan Eichler, Johannes Stefanski, José Martin Roca, Isabelle Daniel

Deep in the Earth's crust, pressure exceeds one thousand times the atmospheric pressure. Water still flows under these conditions, but experiences dramatic changes in structure and fluidity. Using combined dynamic and inelastic light scattering techniques, we simultaneously measure the shear and bulk viscosities of water as a function of pressure. The former

D. J. Manuge

The purpose of this paper is to describe and extend the use of the newly-introduced measure, residual estimation risk. Following the seminal work of Bignozzi and Tsanakas, the quantification of residual estimation risk is proposed in a multivariate framework. Our aim is to provide a succinct and practical introduction to the concept, to motivate its use as a

A. H. Mueller

The size of gluon occupancies, or equivalently the nuclear gluon TMD, at gluon transverse momentum $k_\perp \le Q_s(Y)$ is evaluated. Without Sudakov corrections the occupations can become arbitrarily large while Sudakov effects lead to maximum occupancies of size $(1/\alpha)^{3/2}$. Results are the same for running coupling and fixed coupling dynamics. The

Narjes Ansari, César Feniou, Nicolaï Gouraud, Daniele Loco

Integrating quantum mechanics into drug discovery marks a decisive shift from empirical trial-and-error toward quantitative precision. However, the prohibitive cost of ab initio molecular dynamics has historically forced a compromise between chemical accuracy and computational scalability. This paper identifies the convergence of High-Performance Computing (

Gunnar Brinkmann, Steven Van Overberghe

The essential requirement for a cubic graph to be called a snark is that it can not be edge-coloured with three colours. To avoid trivial cases, varying restrictions on the connectivity are imposed. Snarks are not only interesting in themselves, but also a valuable test field for conjectures about graphs that are not snarks and sometimes not even cubic. For

Finn L. Temmen, Martina Gisti, David J. Luitz, Thomas Luu

Strongly correlated fermionic systems are of great interest in condensed matter physics and numerical methods are indispensable tools for their study. However, existing approaches such as exact diagonalization (ED) and stochastic quantum Monte Carlo methods each suffer from fundamental limitations: ED is hindered by exponential scaling in system size, while

Hamed Taheri

Enterprise AI deploys dozens of autonomous agent nodes across workflows, each acting on the same entities with no shared memory and no common governance. We identify five structural challenges arising from this memory governance gap: memory silos across agent workflows; governance fragmentation across teams and tools; unstructured memories unusable by downst

Elynn Chen, Xi Chen, Yi Zhang

We study transfer learning for contextual joint assortment-pricing under a multinomial logit choice model with bandit feedback. A seller operates across multiple related markets and observes only posted prices and realized purchases. While data from source markets can accelerate learning in a target market, cross-market differences in customer preferences ma

Leonardo Del Grande, Christoph Brune, Marcello Carioni

In this paper, we study total variation (TV)-regularized training of infinite-width shallow ReLU neural networks, formulated as a convex optimization problem over measures on the unit sphere. Our approach leverages the duality theory of TV-regularized optimization problems to establish rigorous guarantees on the sparsity of the solutions to the training prob

Markus Miethlinger, Riccardo Castellano, Pavel Sekatski, Nicolas Brunner

Characterizing the relation between entanglement and Bell nonlocality is a long-standing open problem, notably challenging in the multipartite case. Here we investigate the effect of superactivation of genuine multipartite nonlocality. Specifically, we show that starting from multipartite states that feature only two-party entanglement (hence almost fully se

Haiyu Yang, Sumit Sharma, Enhong Liu, Miel Hostens

Automated behavior classification is essential for precision livestock farming but faces challenges of high computational costs and limited labeled data. This study systematically compared three approaches: training from scratch (ResNet-18, ViT-Small), frozen feature extraction, and parameter-efficient fine-tuning (PEFT) of the DINOv3 foundation model (6.7 b

Oliver Zahn, Simran Chana

Large language models increasingly serve as persistent knowledge workers, with in-context memory - facts stored in the prompt - as the default strategy. We benchmark in-context memory against Knowledge Objects (KOs), discrete hash-addressed tuples with O(1) retrieval. Within the context window, Claude Sonnet 4.5 achieves 100% exact-match accuracy from 10 to

Heng Zhou, Xiaoxiong Liu, Zhenxi Zhang, Jieheng Yun

Remote sensing images (RSIs) are frequently degraded by haze, fog, and thin clouds, which obscure surface reflectance and hinder downstream applications. This study presents the first systematic and unified survey of RSIs dehazing, integrating methodological evolution, benchmark assessment, and physical consistency analysis. We categorize existing approaches

Yizheng Song, Yiyu Zhuang, Qipeng Xu, Haixiang Wang

Single-view 3D human reconstruction has garnered significant attention in recent years. Despite numerous advancements, prior research has concentrated on reconstructing 3D models from clear, close-up images of individual subjects, often yielding subpar results in the more prevalent multi-person scenarios. Reconstructing 3D human crowd models is a highly intr

Petra Mengistu, Karen L. Masters, Tobias Geron, R. J. Smethurst

There is now clear evidence, from a variety of studies, that galactic bars contribute to and/or accelerate processes which quench galaxies. However, bars have a variety of strengths and pattern speeds, and previous work has suggested that slow and strong bars impact their hosts the most. In this paper, we continue to investigate the impact of bar strength an

Mohammed Rafiq Namiq

Let $Δ_{\mathbf r}=Δ_{\mathbf r}(n_1,\ldots,n_e)$ be the clique complex of a chordal graph with maximal cliques $S_1,\ldots,S_e$ in a leaf order, where $n_m=|S_m|$, $r_m=|S_{m+1}\cap(S_1\cup\cdots\cup S_m)|$, $\mathbf r=(r_1,\ldots,r_{e-1})$, and $N_{\mathbf r}$ is the number of vertices. We determine the graded Betti numbers $β_{i,j}\bigl(\mathbb K[(Δ_{\mat

Teng Pan, Yuchen Yan, Zixuan Wang, Ruiqing Zhang

Label-free reinforcement learning enables large language models to improve reasoning capabilities without ground-truth supervision, typically by treating majority-voted answers as pseudo-labels. However, we identify a critical failure mode: as training maximizes self-consistency, output diversity collapses, causing the model to confidently reinforce systemat

Benjamin Hall, Palash Goiporia, Rich Rines

We present Quantum Depth Compression (QDC), a general compilation framework that utilizes dynamic circuits to reduce arbitrary quantum circuits to depth linear in the number of non-Clifford gates and to grid connectivity without the need for expensive SWAP-networks. The framework consists of pushing Clifford gates to the end of the circuit, resulting in a se

Fiammetta Caccavale, Carina L. Gargalo, Julian Kager, Magdalena Skowyra

The landscape of education is changing rapidly, shaped by emerging pedagogical approaches, technological innovations such as artificial intelligence (AI), and evolving societal expectations, all of which demand thorough evaluation of new educational tools. Although large language models (LLMs) present substantial opportunities especially in Higher Education,

Parikshit De, Abinash Panda, Anup Pramanik

We study a public decision problem in which a finite society selects a public-good level from a closed interval. Agents either have single-peaked preferences or are completely indifferent over the interval; the latter capture abstention or a "none of the above" stance within the decision process. We study this augmented single-peaked domain. On this domain,

Yihong Chen, Zhouchen Lin, Quanming Yao

Attention sinks and massive activations are recurring and closely related phenomena in Transformer models. Existing explanations have largely focused on the forward pass, yet in pre-norm Transformers, large residual-stream norms play only an indirect forward role because sublayers operate on normalized inputs. We study this relationship from the perspective

Riccardo Scarpa, Renato Falomo, Aldo Treves

Modifications to Newtonian dynamics at low accelerations have long been proposed as an alternative to dark matter to explain galaxy rotation curves. More recently, similar corrections have been invoked to interpret anomalies in Cavendish-type laboratory experiments and in the dynamics of wide binary stars, although the latter remain affected by ongoing obser

Matthew Wiesner, Samuele Cornell, Alexander Polok, Lucas Ondel Yang

We propose to model parallel streams of data, such as overlapped speech, using shuffles. Specifically, this paper shows how the shuffle product and partial order finite-state automata (FSAs) can be used for alignment and speaker-attributed transcription of overlapped speech. We train using the total score on these FSAs as a loss function, marginalizing over

Khai Yi Chin, Tingwei Meng, Zhe Chen, Daniel Bassett

We propose a novel algorithm for forming arbitrarily shaped assemblies using decentralized robots. By relying on local interactions, the algorithm ensures there are no unreachable states or gaps in the assembly, which are global properties. The in-assembly robots attract passing-by robots into expanding the assembly via a simple implementation of signaling a

Carter Sale, Melissa N. Stolar, Gaurav Patil, Michael J. Gostelow

Real-time cognitive workload monitoring is crucial in safety-critical environments, yet established measures are intrusive, expensive, or lack temporal resolution. We tested whether facial movement dynamics from a standard webcam could provide a low-cost alternative. Seventy-two participants completed a multitasking simulation (OpenMATB) under varied load wh

Radek Zavorka, Tomas Mikulasek, Josef Vychodil, Jiri Blumenstein

This paper presents results from a vehicle-to-vehicle channel measurement campaign conducted in the millimeter-wave (MMW) frequency bands at center frequencies of 60GHz and 80GHz, each with a bandwidth of 2GHz. The measurements were performed in a dynamic oncoming-vehicle scenario using a time-domain channel sounder with high-resolution data acquisition. Pow

Himadri S Samanta

Automated radiology report generation has gained increasing attention with the rise of deep learning and large language models. However, fully generative approaches often suffer from hallucinations and lack clinical grounding, limiting their reliability in real-world workflows. In this study, we propose a multimodal retrieval-augmented generation (RAG) syste

Yingqing Chen, Anni Li, Christos G. Cassandras, Homayoun Hamedmoghadam

Dynamic pricing is commonly used to regulate congestion in shared service systems. This paper is motivated by the fact that in the presence of users with varying price sensitivity (responsiveness), conventional monotonic pricing can lead to unfair outcomes by disproportionately excluding price-elastic users, particularly under high or uncertain demand. We th